Examples of weka.classifiers.functions.SMO

weka.classifiers.functions.SMO

rch.microsoft.com/\~jplatt/smo.html}, PS = {http://research.microsoft.com/\~jplatt/smo-book.ps.gz}, PDF = {http://research.microsoft.com/\~jplatt/smo-book.pdf} } @article{Keerthi2001, author = {S.S. Keerthi and S.K. Shevade and C. Bhattacharyya and K.R.K. Murthy}, journal = {Neural Computation}, number = {3}, pages = {637-649}, title = {Improvements to Platt's SMO Algorithm for SVM Classifier Design}, volume = {13}, year = {2001}, PS = {http://guppy.mpe.nus.edu.sg/\~mpessk/svm/smo_mod_nc.ps.gz} } @inproceedings{Hastie1998, author = {Trevor Hastie and Robert Tibshirani}, booktitle = {Advances in Neural Information Processing Systems}, editor = {Michael I. Jordan and Michael J. Kearns and Sara A. Solla}, publisher = {MIT Press}, title = {Classification by Pairwise Coupling}, volume = {10}, year = {1998}, PS = {http://www-stat.stanford.edu/\~hastie/Papers/2class.ps} }

Valid options are:

 -D If set, classifier is run in debug mode and may output additional info to the console

 -no-checks Turns off all checks - use with caution! Turning them off assumes that data is purely numeric, doesn't contain any missing values, and has a nominal class. Turning them off also means that no header information will be stored if the machine is linear. Finally, it also assumes that no instance has a weight equal to 0. (default: checks on)

 -C <double> The complexity constant C. (default 1)

 -N Whether to 0=normalize/1=standardize/2=neither. (default 0=normalize)

 -L <double> The tolerance parameter. (default 1.0e-3)

 -P <double> The epsilon for round-off error. (default 1.0e-12)

 -M Fit logistic models to SVM outputs.

 -V <double> The number of folds for the internal cross-validation. (default -1, use training data)

 -W <double> The random number seed. (default 1)

 -K <classname and parameters> The Kernel to use. (default: weka.classifiers.functions.supportVector.PolyKernel)

  Options specific to kernel weka.classifiers.functions.supportVector.PolyKernel:

 -D Enables debugging output (if available) to be printed. (default: off)

 -no-checks Turns off all checks - use with caution! (default: checks on)

 -C <num> The size of the cache (a prime number), 0 for full cache and  -1 to turn it off. (default: 250007)

 -E <num> The Exponent to use. (default: 1.0)

 -L Use lower-order terms. (default: no)

@author Eibe Frank (eibe@cs.waikato.ac.nz) @author Shane Legg (shane@intelligenesis.net) (sparse vector code) @author Stuart Inglis (stuart@reeltwo.com) (sparse vector code) @version $Revision: 6025 $

      // using 5 fold CV to emulate the 80-20 random split of jkms
      Instances train = full.trainCV(5, 0);
      Instances test = full.testCV(5, 0);
      
      // new svm with rbf kernel
      SMO smo = new SMO();
      String[] options = { "-L 1e-15", "-P 1e-15" , "-N 2"};
      smo.setOptions(options);
      smo.setC(1.0); //same as default value for jkms
      RBFKernel rbf = new RBFKernel();
      rbf.setGamma(0.1); //same as default value for jkms
      smo.setKernel(rbf);
      
      smo.buildClassifier(train);
      
      
      Evaluation eval = new Evaluation(train);
      eval.evaluateModel(smo, test);

View Full Code Here

   * @see               Capabilities
   */
  public Capabilities getCapabilities() {
    Capabilities        result;
    
    result = new SMO().getCapabilities();
    
    result.setOwner(this);
    
    // only binary attributes are allowed, otherwise the NominalToBinary
    // filter inside SMO will increase the number of attributes which in turn

View Full Code Here

        } else {
          numToElim = (numAttrLeft >= m_numToEliminate) ? m_numToEliminate : numAttrLeft;
        }
        
        // Build the linear SVM with default parameters
        SMO smo = new SMO();
                                
        // SMO seems to get stuck if data not normalised when few attributes remain
        // smo.setNormalizeData(numAttrLeft < 40);
        smo.setFilterType(new SelectedTag(m_smoFilterType, SMO.TAGS_FILTER));
        smo.setEpsilon(m_smoPParameter);
        smo.setToleranceParameter(m_smoTParameter);
        smo.setC(m_smoCParameter);
        smo.buildClassifier(trainCopy);
                                
        // Find the attribute with maximum weight^2
        double[] weightsSparse = smo.sparseWeights()[0][1];
        int[] indicesSparse = smo.sparseIndices()[0][1];
        double[] weights = new double[trainCopy.numAttributes()];
        for (int j = 0; j < weightsSparse.length; j++) {
          weights[indicesSparse[j]] = weightsSparse[j] * weightsSparse[j];
        }
        weights[trainCopy.classIndex()] = Double.MAX_VALUE;

View Full Code Here

        case J48:
          J48 j48 = new J48();      
          j48.setOptions(new String[] { "-C", "0.25", "-M", "2" });
          return j48;
        case SMO:
          SMO smo = new SMO();
          smo.setOptions(Utils.splitOptions("-C 1.0 -L 0.001 -P 1.0E-12 -N 0 -V -1 -W 1 -K \"weka.classifiers.functions.supportVector.PolyKernel -C 250007 -E 1.0\""));
          return smo;
        case LOGISTIC:
          Logistic logistic = new Logistic();
          logistic.setOptions(Utils.splitOptions("-R 1.0E-8 -M -1"));
          return logistic;

View Full Code Here

    public List<Entry<String, Double>> getFeatureWeights(ExampleSet trainingExamples, Model model) {
        WekaModel wekaModel = (WekaModel) model;
        Classifier classifier = wekaModel.getClassifier();
        Instances dataFormat = trainingExamples.getInstances();


        SMO smo = getSMO(classifier);


        double[] sparseWeights = smo.sparseWeights()[0][1];
        int[] sparseIndices = smo.sparseIndices()[0][1];


        Map<String, Double> weights = new HashMap<String, Double>();
        for (int i = 0; i < sparseWeights.length; i++) {
            int index = sparseIndices[i];
            double weight = sparseWeights[i];

View Full Code Here

    private SMO getSMO(Classifier classifier) {
        if (classifier instanceof CostSensitiveClassifier) {
            classifier = ((CostSensitiveClassifier) classifier).getClassifier();
        }


        SMO smo = null;
        if (classifier instanceof SMO) {
            smo = (SMO) classifier;
        } else {
            throw new IllegalArgumentException("Classifier was neither SMO or CostSensitiveClassifier(SMO)");
        }

View Full Code Here

    }


    @Override
    public WekaModel train(ExampleSet examples) {
        System.out.println("SMO Options: " + SMO_OPTIONS);
        SMO smo = new SMO();
        try {
            smo.setOptions(Utils.splitOptions(SMO_OPTIONS));
        } catch (Exception ex) {
            System.err.println("Unable to configure SMO.");
            System.err.println("\t" + ex.getMessage());
            return null;
        }


        //Build logistic models if desired
        smo.setBuildLogisticModels(isBuildLogisticModel());


        Classifier classifier = smo;


        if (useCostTraining) {
            CostSensitiveClassifier cost = new CostSensitiveClassifier();

View Full Code Here

    }


    @Override
    public WekaModel train(ExampleSet examples) {
        //These settings aren't terrible
        SMO smo = new SMO();
        RBFKernel rbf = new RBFKernel();
        rbf.setGamma(0.5);
        smo.setKernel(rbf);
        smo.setC(1.5);
        
        //These also work pretty ok
        Logistic log = new Logistic();
        log.setRidge(100);

View Full Code Here

TOP

Related Classes of weka.classifiers.functions.SMO

CrossValWeka

etc.aloe.cscw2013.SMOFeatureWeighting

etc.aloe.cscw2013.TrainingImpl

etc.aloe.oilspill2010.TrainingImpl

weka.attributeSelection.SVMAttributeEval

weka.classifiers.functions.supportVector.SMOset

weka.core.Attribute

weka.core.Capabilities

weka.core.DenseInstance

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.